Big five personality as a predictor of health: Shortening the questionnaire through the elastic net

Abstract

Background: The Big Five personality attributes (i.e. openness, conscientiousness, extraversion, agreeableness, and neuroticism) help to predict health. To predict health, researchers may prefer to use a short version of the Big Five Inventory. Although the psychometric properties of the shortened scales can be highly satisfactory, their use can lead researchers to substantially underestimate the role of personality. The aim of this paper is to demonstrate a method appropriate for shortening the Big Five Inventory without losing predictive performance. Participants and Procedure: The sample comprised 4,678 panel members. The personality traits were measured in 2017 using the Five Factor Model International Personality Item Pool and subjective health was measured in 2018 using the item "How would you describe your health, generally speaking?" While studying the personality-health relationship, the elastic net was compared to a more conventional regression method. Results: While predicting health based on personality, using 14 Big Five Inventory items (R2= .19) resulted in a similar predictive performance as using 50 Big Five Inventory items (R2= .18). Controlled for gender and age, participants experienced lower levels of health when they "often feel blue", are not "relaxed most of the time", and "worry about things." These aspects of neuroticism relate to the lowerorder facets anxiety and depression. Conclusions: When the primary goal of personality assessment is predictive performance, researchers should consider shortening their questionnaire using the method demonstrated in this paper. Shortening of the questionnaire does not have to result in a lower predictive performance.